The long-term goal of this research program is to develop a complete theoretical account of how children acquire access to segmental speech structure when there are no apparent landmarks for that structure in the acoustic signal. Early work on this project examined how children's sensitivity to and weighting of acoustic cues affected the emergence of segmental structure in their speech perception and production. That work revealed that young children weight dynamic spectral cues (formant transitions) especially strongly and certainly more than adults. Because formant transitions are brief sections of longer, continuously changing patterns of vocal-tract resonances those results were interpreted as suggesting that the research focus should move beyond the cue. Building upon the work of others showing that that there are optimal strategies for perceptually organizing the acoustic speech signal, this project currently focuses on how children learn to use global acoustic structure in speech perception and production. Six Specific Aims will address these hypotheses: Aim #1: Global properties of speech are more readily organized over signal stretches that span several syllables, rather than over brief stretches. Aim #2: Coherence Masking Protection for speech signals arises explicitly from being able to recover a sound-generating object. Aim #3: Children with language-learning problems are particularly poor at perceptually organizing the speech signal in a coherent manner. Aim #4: Cognitive and attentional resources are more greatly taxed with degraded than natural speech signals. Aim #5: Children are more hindered than adults in their speech perception by signals with broad formants. Aim #6: Learning to organize speech gestures in production is more difficult with degraded signals as the models. Attaining these aims will broaden our understanding of what is involved in learning a native language. Outcomes will impact our theories regarding language-related disorders of childhood, and influence treatment.